Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Co-Optimizing Distributed Energy Resources under Demand Charges and Bi-Directional Power Flow

Ruixiao Yang, Gulai Shen, Ahmed S. Alahmed, Chuchu Fan

TL;DR

The paper tackles co-optimizing behind-the-meter DERs—flexible demand, renewable DG, and BESS—under a bi-directional NEM tariff with demand charges using a stochastic dynamic program. It proves a threshold structure for the optimal policy and demonstrates that a simple threshold algorithm can be effective, while reinforcement learning (PPO) closely approaches the optimal benchmark on real-world data. A nonlinear program provides an exact reference, and extensive simulations show substantial surplus gains for PPO over baselines, with the threshold method performing well under certain conditions. The work offers practical guidance for designing prosumer energy management systems under complex tariffs by leveraging policy structure and data-driven learning.

Abstract

We address the co-optimization of behind-the-meter (BTM) distributed energy resources (DER), including flexible demands, renewable distributed generation (DG), and battery energy storage systems (BESS) under net energy metering (NEM) frameworks with demand charges. We formulate the problem as a stochastic dynamic program that accounts for renewable generation uncertainty and operational surplus maximization. Our theoretical analysis reveals that the optimal policy follows a threshold structure. Finally, we show that even a simple algorithm leveraging this threshold structure performs well in simulation, emphasizing its importance in developing near-optimal algorithms. These findings provide crucial insights for implementing prosumer energy management systems under complex tariff structures.

Co-Optimizing Distributed Energy Resources under Demand Charges and Bi-Directional Power Flow

TL;DR

The paper tackles co-optimizing behind-the-meter DERs—flexible demand, renewable DG, and BESS—under a bi-directional NEM tariff with demand charges using a stochastic dynamic program. It proves a threshold structure for the optimal policy and demonstrates that a simple threshold algorithm can be effective, while reinforcement learning (PPO) closely approaches the optimal benchmark on real-world data. A nonlinear program provides an exact reference, and extensive simulations show substantial surplus gains for PPO over baselines, with the threshold method performing well under certain conditions. The work offers practical guidance for designing prosumer energy management systems under complex tariffs by leveraging policy structure and data-driven learning.

Abstract

We address the co-optimization of behind-the-meter (BTM) distributed energy resources (DER), including flexible demands, renewable distributed generation (DG), and battery energy storage systems (BESS) under net energy metering (NEM) frameworks with demand charges. We formulate the problem as a stochastic dynamic program that accounts for renewable generation uncertainty and operational surplus maximization. Our theoretical analysis reveals that the optimal policy follows a threshold structure. Finally, we show that even a simple algorithm leveraging this threshold structure performs well in simulation, emphasizing its importance in developing near-optimal algorithms. These findings provide crucial insights for implementing prosumer energy management systems under complex tariff structures.

Paper Structure

This paper contains 17 sections, 4 theorems, 18 equations, 3 figures, 2 tables.

Table of Contents

  1. Background and Literature
  2. Problem Formulation
  3. Prosumer Resources
  4. System Dynamics
  5. NEM X Tariff with Demand Charges
  6. Flexible Demands and Storage Co-Optimization
  7. Dynamic Program Analytical Results
  8. Case Study
  9. Problem Setup
  10. Control Strategies
  11. Baseline
  12. Reinforcement Learning using PPO
  13. Threshold Algorithm
  14. Nonlinear Program
  15. Dataset
  16. ...and 2 more sections

Key Result

Lemma 1

The optimal value function $V^*_t(s_t, g_t, c_t)$ is concave in $(s_t, c_t)$ for every time step $t\le T$.

Figures (3)

  • Figure 1: Distributed energy resource setup: flexible demands $d\in \mathbb R_+$, renewable DG $g\in\mathbb R_+$, storage operation $e\in \mathbb R$, and net-consumption $z\in \mathbb R$.
  • Figure 2: Daily net demand data from June 2022 to May 2024
  • Figure 3: Typical day battery state of charge and demand comparison between different control methods

Theorems & Definitions (8)

  • Lemma 1
  • proof
  • Lemma 2
  • proof
  • Lemma 3
  • proof
  • Theorem 1
  • proof